isolation transformer install

[eschertron]

Nice results! Did you happen to get a shot (or voltage measurement) of the signal at power tube grid? plate? I'd like to know how close you came to Juan's estimated 200vpp on the plate as it clips.

I'll have to dig up a shot of a champ's output when I get home. For me, the waveform you show looks way too "nice" to represent a SE amp in clipping. Of course, I know just shy of less than nothing about the power tube in your amp, so my expectations could be skewed.

[Randall]

I see 140v p-p at the plate.

[eschertron]

This is my experience with SE power amp clipping. Leads to that archetypal raspy timbre.

I'm not doubting your results, but I suspect the signal is being limited before the power tube grid, choking your potential for more power. Just my 'tin-foil-hat' hypothesis of the day. I hope you are happy with the results you are getting - or if not - I hope you get to those results. Every amp should be able to speak with its own voice.

[J M Fahey]

Here is the waveform driven at input with 1KHz at 120mV (measured with Fluke non true RMS), output is measured across a 3.9 ohm 5 watt resistor with mystery OT. Scope is set at 0.5v/div for approx. 2.5v p-p. My calculations put that at .20 watts. 2.5 v p-p / 2 = 1.25 v peak, 1.25v * .707 = 0.88 v rms, (0.88 * 0.88)/ 3.9 = 0.20 watts.
[ATTACH=CONFIG]49870[/ATTACH]

I see 140v p-p at the plate.

Cool, we are getting closer.
Just as an experiment, if you have another 3.9 ohm resistor, let´s call it 4 ohm for practicity, re measure amp with 8 ohm (both in series) and 2 ohm (both in parallel) so in a rough way we know better what impedance it likes more.
Again post pictures because not necessarily it will clipthe same way, with different impedance maybe bottom clips before top ... or both become more symmetrical, who knows without testing?

Hopefully one of those doubled or halved loads provides somewhat more power.

And after that we might tweak idle bias, which could also yield a bit more.

I am asking you just what I would experiment on my own bench

[Randall]

I don't have another 3.9 ohm, but I do have a 5 ohm 5 watt, can we work with that?

[Enzo]

I may have missed it, but are we using a X10 probe?

[Randall]

"I may have missed it, but are we using a X10 probe?"

No, it's not switched to x10, I verified that more than once.

[Leo_Gnardo]

I don't have another 3.9 ohm, but I do have a 5 ohm 5 watt, can we work with that?

Sure, give it a shot. Whichever way you measure more power you know you're heading in the right direction. Good suggestion Juan on trying series & parallel combinations.

[Randall]

I have an 8 ohm and a 5 ohm resistor, and here is what I see with 120mV input, 0.5v/div

8 ohm, a little over 2.5 v p-p


2.4 ohm, about 2 v p-p


I don't know what to make of this?

[uneumann]

I have an 8 ohm and a 5 ohm resistor, and here is what I see with 120mV input, 0.5v/div

8 ohm, a little over 2.5 v p-p
[ATTACH=CONFIG]49880[/ATTACH]

2.4 ohm, about 2 v p-p
[ATTACH=CONFIG]49881[/ATTACH]

I don't know what to make of this?

Just a wild guess ... have you replaced or put a cap in parallel with C5A - the bypass cap?
If that's not truly bypassing the bias resistor R7 you'll lose some power there without impacting your grid and plate voltage measures.

[Jazz P Bass]

Is it me?
I am so used to dealing in RMS voltage measurements that I can't wrap my head around "p-p"

Will the output formula "V squared divided by resistance" still apply? (admittedly crudely)

[Enzo]

Cut p-p in half for peak, then 0.707 of that for RMS.

[Randall]

"Just a wild guess ... have you replaced or put a cap in parallel with C5A - the bypass cap?
If that's not truly bypassing the bias resistor R7 you'll lose some power there without impacting your grid and plate voltage measures."

yes, there is a new 25/25v in there now.

[uneumann]

"Just a wild guess ... have you replaced or put a cap in parallel with C5A - the bypass cap?
If that's not truly bypassing the bias resistor R7 you'll lose some power there without impacting your grid and plate voltage measures."

yes, there is a new 25/25v in there now.

Hmmm - 25uf is common for a 1.5K bias resistor, but you have a 150 ohm resistor. I compute Xc = ~32 ohms at 200Hz for a 25uf.
You may want to increase the bypass cap by 10x to 250uf or more. If may not make a huge difference, but you're looking for any scraps of power you can get.

[Randall]

I'm not trying to re-engineer this thing. I just want to know if it is working as designed. So far, it doesn't seem so. Or someone prove me wrong. Please.